Abstract:Quantum machine learning has the potential to enable advances in artificial intelligence, such as solving problems intractable on classical computers. Some fundamental ideas behind quantum machine learning are similar to kernel methods in classical machine learning. Both process information by mapping it into high-dimensional vector spaces without explicitly calculating their numerical values. We explore a setup for performing classification on labeled classical datasets, consisting of a classical neural network connected to a quantum annealer. The neural network programs the quantum annealer's controls and thereby maps the annealer's initial states into new states in the Hilbert space. The neural network's parameters are optimized to maximize the distance of states corresponding to inputs from different classes and minimize the distance between quantum states corresponding to the same class. Recent literature showed that at least some of the "learning" is due to the quantum annealer, connecting a small linear network to a quantum annealer and using it to learn small and linearly inseparable datasets. In this study, we consider a similar but not quite the same case, where a classical fully-fledged neural network is connected with a small quantum annealer. In such a setting, the fully-fledged classical neural-network already has built-in nonlinearity and learning power, and can already handle the classification problem alone, we want to see whether an additional quantum layer could boost its performance. We simulate this system to learn several common datasets, including those for image and sound recognition. We conclude that adding a small quantum annealer does not provide a significant benefit over just using a regular (nonlinear) classical neural network.

Neural networks with quantum states of light

Shedding Light on the Future: Exploring Quantum Neural Networks through Optics

Realistic quantum photonic neural networks

Quantum Algorithms for Feedforward Neural Networks

Complex Deep Learning with Quantum Optics

Quantum Information and Artificial Neural Networks-based Signal Processing Using Integrated Photonics

Hybrid Quantum-Classical Photonic Neural Networks

Quantum optical neural networks with programmable nonlinearities

Quantum consistent neural/tensor networks for photonic circuits with strongly/weakly entangled states

Neural-Network Quantum States: from Condensed Matter to Quantum Computing

First Photon Machine Learning

Quantum machine learning via continuous-variable cluster states and teleportation

Quantum Neural Network for Quantum Neural Computing

Quantum Computing and Machine Learning on an Integrated Photonics Platform

A Quantum Optical Recurrent Neural Network for Online Processing of Quantum Times Series

A Review of the Applications of Quantum Machine Learning in Optical Communication Systems

Quantum-dot based photonic quantum networks

Neural Networks for Programming Quantum Annealers

Lensless Opto-Electronic Neural Network with Quantum Dot Nonlinear Activation

A Quantum Hopfield Neural Network Model and Image Recognition